The Study On Computer Dynamic Simulation Of Maize Growth

This research took maize as object, based on crop physiology, ecology theories and experiment data, used the system analysis theories and the mathematical modeling technology to study the knowledge representation system of maize cultivation management. This research looked maize production as a total system consisted of crop, environment, technology, economy and so on. It synthetically used computer simulation, software theory and digitalization technology, studied maize production target, moisture requirement, information acquisition, information procession, information management and information application etc pivotal technology. So it digitalized the information stream of maize growth, made moisture requirement and production information from qualitative analysis to quantitative analysis, from concept model to digital model, form expert experiment to optimization decision-making. It made a basic for time, quantitative and pinpoint digitalization and accurate crop information management and decision support. It analyzed the light productivity, the temperature moisture management and the nutrient supply, cultivation management level etc. it established a maize production target model. According to the field potential moisture evaporation, the effective rainfall, it established a maize moisture requirement dynamic model. The paper studied the model solution, used variable describe the system state, used math function describe the correlation of variables and translated the concept model into quantitative math model of the question solution. It designed a simulation software based on B/S structure and implemented the computer simulation of maize production and moisture requirement. It implemented the digital design of maize production project and management adjustment target. It optimized the model, used the floating point numbers coding to take the place of binary coding and introduced the population competition mode, used an adaptive adjustment evolution algorithm of an adaptive value based on population classification. The algorithm took the kin individuals as a population, at the initial stages, the evolution searching only in the population inside so that individuals' evolution can not be effect by the other populations and every population had chances to produce excellent individuals. After evolutionary computing several generations, this algorithm let all of the individuals compete and evolve each other so that the worse individuals gradually dying out of the individual competition. This algorithm fully maintained the diversity of populations, effectively conquered the premature convergence and local optimum solution. It described evolution operators in detail. For the importance of variation operation in floating-point number coding, the paper used self-adapting adjusting variation operator, adjusted the variation steps based on individual self-adapting value so that the better individuals varied in little steps in local region, enhanced the precision, the worse individuals increased the steps to get out of the worse regions and enter into the prospective regions.